A modular platform for fusion of images

Claire Thomas (1) , Jean-Christophe Laneri (2) , Thierry Ranchin (1) , Lucien Wald (3)

1 CEP - Centre Énergétique et Procédés
2 ESIEA - Ecole Supérieure d'Informatique Electronique Automatique [Paris]
3 CEP/Sophia

This paper presents a platform for the assessment of strategies for the fusion of images of various modalities and having different spatial resolution. The purpose of the fusion is the synthesis of multimodal images offering the best spatial resolution available in the data set. These synthesized images should be as close as possible to reality. This platform is written in IDL. It calls upon the ARSIS concept and thus comprises three categories of models. For each category, several models were implemented. Each model of a given category may be combined with models of the other categories, thus offering several possible strategies. The part « quality assessment » provides quantitative values for the fusion. It can also be executed, independently from the fusion part, on images resulting from a fusion process outside the platform. The platform should be made public.

Domaines: Autre; Traitement du signal et de l'image [eess.SP]; Sciences de l'environnement; Sciences de la Terre; Traitement du signal et de l'image [eess.SP]

Claire Thomas, Jean-Christophe Laneri, Thierry Ranchin, Lucien Wald. A modular platform for fusion of images. 4th International Conference on Physics in Signal Image Processing PSIP 2005, Toulouse, France, January 3rd - February 2nd, 2005, Jan 2005, Toulouse, France. p. 269-273 - ISBN 2-912328-22-5. ⟨hal-00068152⟩